The problem it solves
Business stakeholders wait days for an analyst to answer a question that's really just a differently-phrased version of an existing report. Self-service BI tools help, but still require knowing which dashboard to open and how to filter it correctly.
Conversational Analytics collapses that gap: ask the question directly, get an answer grounded in your actual governed metrics and data, with the underlying query and source shown for verification — not a black box.
How it works
Natural-language questions are translated into queries against your existing data warehouse or lakehouse, executed with row- and column-level security enforced at the query engine for the asking user's actual identity, exactly as our data engineering practice recommends for any LLM-to-data integration.
Every answer links back to the underlying query and source tables, and to your existing metric registry where one exists, so an analyst can verify or drill deeper without starting from scratch.
What's included
Implementation against your existing warehouse or lakehouse (Snowflake, BigQuery, Redshift, Databricks and open table formats including Apache Iceberg are supported), permission-model integration with your identity provider, and a pilot scoped to a specific business question set before any wider rollout.
This is offered as a scoped professional services engagement combined with the underlying product — not an off-the-shelf SaaS signup — because the permission-model integration genuinely needs to match your organisation's actual access model.
A model that answers nicely is not the same as a model that answered only what the user was allowed to see — that discipline is built into the product, not bolted on after.MNB Technology
What this guarantees
- Enforce row/column security at the query engine for every question
- Link every answer back to its source query and tables
- Integrate with your existing metric registry where one exists
- Start with a scoped pilot before organisation-wide rollout